By Monica Heger
This story was originally published Feb. 16.
Amid building anticipation for the commercialization of a nanopore sequencing system, Mark Akeson's group from the University of California, Santa Cruz, has demonstrated an automated system for "ratcheting" DNA back and forth through a nanopore at a slow enough speed to enable the detection of individual bases.
The system, which makes use of the processive enzyme phi29 DNA polymerase to slow DNA as it races through the pore, was described online last week in Nature Biotechnology.
Akeson's team originally demonstrated the ratcheting method two years ago, showing that DNA translocation can be controlled by positioning a DNA polymerase at the edge of a nanopore. The polymerase initiates replication of a single strand of DNA, moving it through the nanopore only when the strand is captured in the pore and a certain voltage is applied (IS 9/28/2010).
The problem with the original method was that the polymerase "did not hang onto the DNA very well," Akeson explained. In a subsequent paper, his team demonstrated that phi29 held onto the DNA more efficiently.
Now, in its latest paper, his team has developed a "fully automated process to pull the DNA through by unzipping" it. As soon as it goes through the pore in one direction, the polymerase initiates replication, driving it back through the pore in the other direction, getting both the forward and reverse reads, Akeson said.
The basic concept could be applied to any nanopore sequencing system, such as the strand sequencing system that Oxford Nanopore has said it will commercialize this year (IS 2/7/2012).
Oxford Nanopore declined to disclose whether it plans to incorporate elements of Akeson's system in the commercial version of its system, although it does have an exclusive license to some of the technology being developed by his lab.
In an e-mail, Gordon Sanghera, Oxford's CEO, said that the company will talk more about "the metrics of our custom nanopore machinery" at the Advances in Genome Biology and Technology meeting this week in Marco Island, Fla.
A spokesperson added that the company does not plan to disclose the precise nanopore and enzyme that it will use in its system and noted that the metrics of its system are different than Akeson's, which was designed in an academic setting.
Blocking Oligomer
Six main features are required for a working nanopore sequencing system: the automated capture and processing of DNA templates, spatial control over the DNA so that only one base moves through the pore at a time, temporal control over the DNA so that it moves through the pore slow enough to recognize bases, the absence of a complex voltage system that would cause crosstalk in an array of thousands of nanopores, a sensor that can resolve bases, and a counter that can identify nucleotide transistions in homopolymeric regions.
"This work satisfies four of the six criteria needed for nanopore sequencing and is a general model that others could improve upon," Akeson said.
One of the key factors to automating his system was the introduction of a blocking oligomer. If all the ingredients for DNA replication are placed in solution above a nanopore, "before you could catch the molecule and read it, all the DNA would be replicated," he said. "So you have to have some way to prevent the polymerase from replicating in solution and only activate at the nanopore."
In the study, the team tested a 70-nucleotide synthetic DNA template with a 23-nucletoide primer attached to its 5' end, followed by a 25-nucleotide blocking oligomer, which was complementary to the template strand. Appended to the blocking oligomer were seven abasic residues that served to protect the blocking oligomer from exonucleolysis by phi29 and to help remove the blocking oligomer as the whole complex is pulled into the nanopore by an applied voltage.
The team next demonstrated that applying voltage drives the DNA complex into the nanopore, which initiates phi29 to begin "unzipping" the blocking oligomer from the DNA template, ratcheting the template through the pore. Once the blocking oligomer is removed and the template has been driven through the pore, the 3' end of the template is captured within the polymerase, which serves to initiate replication, and the template is then driven back through the pore in the reverse direction.
Next, the group shortened the length of the blocking oligomer to 15-nucleotides to increase throughput of the system. They demonstrated that 500 DNA molecules could be processed through one pore at a rate of about 130 molecules per hour. The DNA strands were ratcheted through the pore at rates of 2.5 nucleotides to 40 nucleotides per second.
What's important about the system from a commercial point of view is that "because of the blocking oligomer, the DNA polymerase will bind to the DNA in solution and only be activated when you pull on it with the nanopore," said Akeson. "So, you can put everything you need in a sequencing apparatus and let it run for hours and hours."
The team next estimated the error rate of the system, focusing on two types of errors — one in which the DNA molecule moves through the pore too quickly for the sensor to detect it, which would result in a deletion; and another in which the strand slips back and forth so that a particular base is read more than once, which would result in an insertion. By assessing the ionic current states that corresponded to each of these patterns, the researchers predicted that the probability of an indel registry error at a given position during one pass along the template ranged from 10 percent to 24 percent.
Akeson noted, however, that the system has not yet been optimized, and to a certain extent, improving the error rate would be a straightforward process.
One way to improve on the error rate, he said, would be to study other enzymes that may be more efficient than phi29. Another way, he said would be to change the "ionic strength." The current through the nanopore is dependent on the concentration of salt, Akeson explained. In the Nature Biotechnology study, the researchers had a .3 molar salt concentration, but increasing that to 1 molar or 2 molar would improve the signal-to-noise ratio, he said.
Additionally, the pore itself could be improved. Akeson's team used an alpha hemolysin protein pore in this study, but he has also been collaborating with Jens Gundlach at the University of Washington, who uses Mycobacterium smegmatis porin A, or MspA, as a protein pore (IS 3/29/2011).
Gundlach previously demonstrated that the MspA may be superior to alpha hemolysin because it is smaller and simpler, so the signal-to-noise ratio is lower (IS 8/24/2010).
Nevertheless, Akeson said there are still a number of other hurdles that need to be addressed. Among other improvements, his team is now working on developing a "clock" for counting bases as they pass through the pore, which would be particularly important in homopolymeric regions.
Have topics you'd like to see covered by In Sequence? Contact the editor at mheger [at] genomeweb [.] com.